Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: MichaelKoerner on May 04, 2020, 03:38:11 pm
-
A couple of months ago I started writing profiles für my Epson SC-P800 with i1Profiler/i1Pro2.
Now, comparing my own profile for Hahnemühle MuseumEtching with the "canned" profile provided by Hahnemühle, I see:
- At Digitaldog's Printer Gamut Test File: With Perceptual R.I., canned profile shows visually darker primaries and severe banding issues in Bill's balls and gradients. Mine is much smoother. Rather the same with Rel.Col.
- Using Rel.col., Test images without Bill's balls show marginal visual difference between the two profiles.
- Using Perceptual R.I. on such images, the "canned" profile seems to better balanced, with better rendering of dark shadows.
As a beginner, on a first look at Andrew's Gamut Test File, I'd say my profile is better then Hahnemühle's (regarding transitions, overall tonality, three-dimensionality). I assume it is "more colorimetrically correct".
But I ask myself (and subsequently, you ;-)
- whether I missed some point. Has the "canned" profile perhaps some hidden advantages I haven't discovered yet?
- Can those banding issues be the side effect of "intentional" profile tuning?
- Are those banding issues practically relevant? What explanatory power do they have?
Of course, I could settle on "use what pleases". But at this my stage of the learning curve, I would prefer some analysis to gain better understanding of the fundamentals of color management.
Please find attached a screenshot of the compared softproofs (Perc. RI), a scanned print of the compared printing results (Perc. RI) and the profiles themselves. I wrote mine (the one starting with MK_) with 1230 patches, 2880dpi, bidirectional. I measured the printed target twice and let i1Profiler average the data.
Any suggestions on how to fulfill more compelling tests to evaluate and compare profiles (those and in general) are very welcome!
Regards, Michael
PS: Haven't overpowered myself to purchase ColorThinkPro yet. Same for PatchTool. But if such analysis required some of those tools, I would go for it.
-
There are no rules in how a profile maker creates a Perceptual rendering. It is up to each manufacturer to decide what they think their customers will prefer. Like E6 film.
One profile has far more visible banding than the other. Not good.
-
One profile has far more visible banding than the other. Not good.
Andrew, thanks for your reply.
As this is the profile provided by Hahnemühle, I wonder whether this is on purpose or just a bad profile.
-
It's not on purpose, it's not necessarily a 'bad' profile but you've done better.
-
Check www.colorcheck-online.com, it has a good test approach, especially for less saturated colors like skin color. It is just one test, use it in addition to the testfile of Andrew, as that test will show in particular the quality at the boundary of the profile, I.e. how it handles out of gamut colors.
A couple of months ago I started writing profiles für my Epson SC-P800 with i1Profiler/i1Pro2.
Now, comparing my own profile for Hahnemühle MuseumEtching with the "canned" profile provided by Hahnemühle, I see:
- At Digitaldog's Printer Gamut Test File: With Perceptual R.I., canned profile shows visually darker primaries and severe banding issues in Bill's balls and gradients. Mine is much smoother. Rather the same with Rel.Col.
- Using Rel.col., Test images without Bill's balls show marginal visual difference between the two profiles.
- Using Perceptual R.I. on such images, the "canned" profile seems to better balanced, with better rendering of dark shadows.
As a beginner, on a first look at Andrew's Gamut Test File, I'd say my profile is better then Hahnemühle's (regarding transitions, overall tonality, three-dimensionality). I assume it is "more colorimetrically correct".
But I ask myself (and subsequently, you ;-)
- whether I missed some point. Has the "canned" profile perhaps some hidden advantages I haven't discovered yet?
- Can those banding issues be the side effect of "intentional" profile tuning?
- Are those banding issues practically relevant? What explanatory power do they have?
Of course, I could settle on "use what pleases". But at this my stage of the learning curve, I would prefer some analysis to gain better understanding of the fundamentals of color management.
Please find attached a screenshot of the compared softproofs (Perc. RI), a scanned print of the compared printing results (Perc. RI) and the profiles themselves. I wrote mine (the one starting with MK_) with 1230 patches, 2880dpi, bidirectional. I measured the printed target twice and let i1Profiler average the data.
Any suggestions on how to fulfill more compelling tests to evaluate and compare profiles (those and in general) are very welcome!
Regards, Michael
PS: Haven't overpowered myself to purchase ColorThinkPro yet. Same for PatchTool. But if such analysis required some of those tools, I would go for it.
-
Jan, thank you for this very interesting link. Much appreciated.
-
Check www.colorcheck-online.com
Jan, this website (.com) asks me for username/password instantly. Different TLDs (.net, .org) seem to mirror to this same site.
Care to share login data, if generally accessible?
In case you wanted to point me to de german TLD (.de, here one can get evaluations of uploaded color control strip measurements) - thanks, but I could do that in EFI Verifier.
I was just curious about comparing/analysing ICC profiles using software. If someone would point me a resonably path to go, I'd be very glad. Until then ... I read Color Think Pro manual ;-)
-
Jan, this website (.com) asks me for username/password instantly. Different TLDs (.net, .org) seem to mirror to this same site.
Care to share login data, if generally accessible?
In case you wanted to point me to de german TLD (.de, here one can get evaluations of uploaded color control strip measurements) - thanks, but I could do that in EFI Verifier.
I was just curious about comparing/analysing ICC profiles using software. If someone would point me a resonably path to go, I'd be very glad. Until then ... I read Color Think Pro manual ;-)
How do you want to analyse profiles? Visual, numerical ?
I use colorcheck-online (tells allot about non saturated colors), my own 51step linear Lab grey wedge (tells about the white to black transition and also quick acid test if something has changed), copra ( colorlogic.de , the profile manager part is public) for viewing profiles & rudimentary comparing, CTO for more in depth analysis, excel with dE 2000 formulas, and visual tests with certain testprints. Remember that what you see and what you measure are not necessarily interchangeable, but more complementary.
My visual tests are still cornerstone.
-
Remember that what you see and what you measure are not necessarily interchangeable, but more complementary.
Yes and I suppose it's worth pasting this again in this thread:
Colorimetry and the dE testing and the creation of profiles is about (based on) color perception. It is not about color appearance. There's a significant difference. The reason why viewing a print is more valid than measuring it is because measurement is about comparing solid colors. Color appearance is about evaluating images and color in context which measurement devices can't provide. Colorimetry was never designed as a color appearance model. It was never designed to even be used as an interchange space between device dependent color models. It's not designed for imagery at all. Colorimetry based on solid colors in very specific ambient and surround conditions.
-
How do you want to analyse profiles? Visual, numerical ?
In this case, numerical. Because,
- Museum Edge is (for me) a very expensive paper. If I do printing tests, they should be as efficient (aka structured) as possible.
- Visual comparing with Hahnemühle's own profile showed differences I want to understand.
- I want to learn about the "behind the scenes"
.
I use colorcheck-online (tells allot about non saturated colors),
Care to answer my question about the correct top level domain?
copra ( colorlogic.de , the profile manager part is public) for viewing profiles & rudimentary comparing, CTO for more in depth analysis
Thanks for the hint to CoPrA (was not on my radar yet).
Remember that what you see and what you measure are not necessarily interchangeable, but more complementary.
My visual tests are still cornerstone.
Clear on that (also thanks, @Andrew, for your explanation). As said, this is mainly out of curiosity (and austerity ;) )
-
Thanks Andrew, for making clear why.
Well worth posting why, and specifically the sobering of measurements.
Yes and I suppose it's worth pasting this again in this thread:
Colorimetry and the dE testing and the creation of profiles is about (based on) color perception. It is not about color appearance. There's a significant difference. The reason why viewing a print is more valid than measuring it is because measurement is about comparing solid colors. Color appearance is about evaluating images and color in context which measurement devices can't provide. Colorimetry was never designed as a color appearance model. It was never designed to even be used as an interchange space between device dependent color models. It's not designed for imagery at all. Colorimetry based on solid colors in very specific ambient and surround conditions.
-
Jan, Andrew,
I'd love to hear your confirmation whether I understand you aright:
- Looking at a profile itself (ColorThink Pro, Copra's Profile Manager) tells not the whole story
- To evaluate this profile's performance you print
- test images (like Andrew's Printer Gamut Test) for visual judgement
- test patches using that new profile to measure the results and analyze them (for example at color-checker.de).
- In general, you rank visual results higher then colorimetrical ones, using the latter as a supplement.
Did I get that right? Thanks in advance for your time and effort helping me sorting things out.
Regards, Michael
-
I use colorcheck-online (tells allot about non saturated colors),
Care to answer my question about the correct top level domain?
The side is dead that's why you asked for login details. They were selling their software for 15000Eur, guess nobody was foolish to buy it after all many bugs were in that software.
-
Really?
https://www.colorcheck-online.de/en.html
The side is dead that's why you asked for login details. They were selling their software for 15000Eur, guess nobody was foolish to buy it after all many bugs were in that software.